home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.fortran
- Path: sparky!uunet!cs.utexas.edu!hermes.chpc.utexas.edu!michael
- From: michael@chpc.utexas.edu (Michael Lemke)
- Subject: Re: A VMS Fortran I/O question
- Message-ID: <1993Jan5.061934.15535@chpc.utexas.edu>
- Organization: The University of Texas System - CHPC
- References: <MARBY.93Jan2161438@laura.harvard.edu> <1993Jan3.040453.20688@chpc.utexas.edu> <C0CGGB.1tE@news.cso.uiuc.edu>
- Date: Tue, 5 Jan 93 06:19:34 GMT
- Lines: 54
-
- In article <C0CGGB.1tE@news.cso.uiuc.edu> ercolessi@uimrl3.mrl.uiuc.edu (furio ercolessi) writes:
- >In article <1993Jan3.040453.20688@chpc.utexas.edu>, michael@chpc.utexas.edu (Michael Lemke) writes:
- >|> I'd rather think the *Unix* stuff should honor
- >|>FORTRAN carriage control. Well, I know it doesn't but a PRINT
- >|>statement after all creates a print file (or even sends something to a
- >|>printer), so doesn't FORTRAN require print files to honor the carriage
- >|>control character? What is the excuse this time that the Unix joke is
- >|>still standard conform?
- >
- >They have put there a filter called asa, and you just pipe the output
- >through it. But then you will have a bit of a problem for interactive
- >programs (buffering will be your enemy!), and also for "batch"
- >programs writing on a file if you like to peek at the output while the
- >program is running.
-
- Yes, I know that beggar and associated problems. In my opinion this is
- not a solution. With the same argumentation they could output troff or
- TeX or postscript. Doesn't matter if the filter is called asa or
- nroff. In either case you get a file which does not print propoerly
- without further action. I am still waiting for an explanation if this
- Unix behavior is standard conform or not (sorry, I don't have a copy of
- the standard to check myself).
-
- >
- >|> BTW, in Sun Fortran there is a special STATUS='PRINT' in
- >|>the OPEN statement which causes the first character to be ignored if a
- >|>blank or to be converted to a CTRL/L if a 1 etc. `+' does not work
- >|>though. This evne works for stdout but I don't know about PRINT. Maybe
- >|>your Unix compiler has something similar.
- >
- >This sounds a neat idea and I wholeheartly suggest that all Unix compiler
- >writers put this feature in (hopefully, with the same syntax ...).
- >I only have two questions: how can you OPEN the unit where the PRINTs
- >go (usually mapped to the Unix standard output) ? OPEN(*,) ???
- >And, are you sure that the keyword is 'STATUS'? I think that STATUS
- >has another meaning associated with it and cannot be overloaded.
-
- Yes, they do overload STATUS and this is why I don't consider it a nice
- idea. Also, I think all files should have this attribute set by
- default (see above) and second, I wonder if the standard says anything
- about the external (file) representation of the carriage control
- character -- does it have to be in the file literally? If not,
- something like the Sun solution would really be an acceptable way to
- deal with it under Unix. asa et al. certainly is not.
-
- I don't know right now how you can affect `channel *' or PRINTs. PRINTs
- might have a negative channel number or something else special. * might
- not work at all.
-
- Michael
- --
- Michael Lemke
- Astronomy, UT Austin, Texas
- (michael@io.as.utexas.edu or UTSPAN::UTADNX::IO::MICHAEL [SPAN])
-